Yankee Group issued its latest survey findings on SOA adoption among 300 US companies, concluding that "2006 will be a benchmark year" for SOA. "What once was used primarily by large enterprises to rescue applications and processes in distress is now reaching ubiquitous implementation in the US, regardless of size or vertical industry designation," the analyst firm says in its announcement of the results.
Yankee Group reports that SOA retail will show the fastest adoption rate in the next 12 months;, planned use of SOA in 2006 reaches saturation for wireless (93%), retail (92%), financial (89%), manufacturing (76%) and government (75%). Yankee predicts that "2006 will be the year of initial SOA project completion on a broad basis--not a hit or miss trend, but through a rising tide of broad and deep adoption of SOA across the market."
These results seem too ambitious. I haven't seen the guts of the report, nor what constituted the definition of an SOA in this survey. Certainly, we are starting to see some shining examples of SOA projects that have come to fruition, such as KeyBank, or Fireman's Fund, as Britton has cited in his recent posts.
Other surveys are also unquestionably finding these kind of adoption levels for XML and Web services, but there's a gigantic leap between having a lot of XML/Web services deployments and having a well-orchestrated, enterprise-class, fully functioning SOA. Some of the standards that will actually make SOA work -- such as Business Process Execution Language, or BPEL, and Web Services Distrubuted Management, or WSDM -- have just barely gotten out the gate.
But I think its safe to say that seven, eight, or nine out of ten enterprises will be working toward SOA, but I don't know if all can make it there in 12 months' time. It's still early in the game.